AppLovin (NASDAQ:APP) reported Q2 2026 revenue of $1.92 billion, missing Wall Street's $1.94 billion estimate by 1.2%, though revenue still grew 53% year-over-year. EPS of $3.76 came in line with expectations, and adjusted EBITDA hit $1.61 billion. Shares tanked 18–21% in premarket and post-earnings trading Thursday.
CEO Adam Foroughi attributed the revenue shortfall entirely to timing: deep-learning model improvements—the invisible engine driving incremental advertiser spend—arrived later than usual in Q2, with a significant upgrade launching just after quarter-end. Adjusted EBITDA margin held near 84% despite the miss, signaling operational discipline.
Q3 guidance of $2.055–$2.085 billion (46–48% year-over-year growth) and $1.71–$1.74 billion EBITDA reflected renewed momentum from post-quarter model improvements already live. The consumer advertising vertical (primarily e-commerce) hit an all-time high, with advertiser spend 28% above Q4 2025's seasonal peak—a notable bright spot offsetting gaming segment slippage.
Piper Sandler analyst James Callahan downgraded APP to neutral and slashed the price target from $665 to $385, citing questions about beat/raise cadence. Architects tracking adtech and AI-powered targeting should watch whether Q3 delivers on accelerated model performance and whether margin holds as compute costs rise.